DESCRIPTION: (Verbatim from applicant's abstract) The aim of this competitive renewal is to complete the ongoing, blinded, placebo-controlled, cross-over study of the effectiveness of the ketogenic diet in children with the intractable atonic/myoclonic seizures associated with the Lennox-Gastaut (LGS) syndrome. This renewal will allow us to complete the cohort of 20 patients and the follow-up and analysis of those enrolled. Since the time of our initial submission, the ketogenic diet has acquired increasing acceptance for the treatment of children with difficult-to-control seizures. However, it still has never been studied in a controlled fashion. Our initial grant was designed to evaluate the diet in this fashion. There were difficulties with recruitment due to our requirement that children must have insurance coverage and transportation to Hopkins prior to recruitment for the study. Since we have been able to abolish those requirements, we have dramatically increased our recruitment and our enrollment. We have, to date completed the cross-over studies of 16 children. We anticipate that transition funding will allow us to enroll several additional children. We are re-submitting this competitive renewal in order to complete the recruitment, enrollment and analysis of the remaining patients, permit us to follow each for one year, and permit us to complete the analysis. Our protocol remains unchanged. After a 24 hour EEG has confirmed more than 15 electro-clinical events, the protocol is initiated with a 36 hour fast followed by increasing portions of meals over 3 days. After a repeat 24 hour EEC the child is re-fasted and the diet is re-initiated. At the end of the second initiation, the child has a third 24 hour EEG and is discharged home on the full ketogenic diet. During the initial arm of the study children are randomized to a liquid supplement containing either placebo (maintaining ketosis) or glucose (negating ketosis). During the second phase, the child receives the alternate supplement. Only the study coordinator who is not involved in the child's care is unblinded. After discharge the child is followed by phone in standard clinical fashion with the parent maintaining monthly seizure calendars. Each child returns for follow-up at six months and one year for clinical and laboratory examination and a repeat 24-hour EEG. The primary endpoint of the cross-over study is a comparison of the number of electro-clinical events (seizures) at the end of each epoch of the crossover study to baseline and within the treatment arms. The mechanics of the study have gone well with no adverse events and no dropouts. We believe that the recruitment problems have been solved. Since the investigators remain blinded, the results are not available.